RE: SPARQL Update - Clarifying correct behavior when a graph is created implicitly

From: Stephen Allen <sallen@bbn.com> · Date: Wed, 14 Sep 2011 12:36:52 -0400

Hi David,

This is definitely a bug in Parliament.  It's a remnant from a previous WD that required explicit CREATE operations before insertions were allowed.  I think the current WD intention is fairly clear, but your addition does make it explicit.

-Stephen

P.S.  The mailing list for Parliament is located at [1] and is subscribed to by all of the developers.

[1] http://www.semwebcentral.org/mailman/listinfo/parliament-users

> -----Original Message-----
> From: public-rdf-dawg-comments-request@w3.org [mailto:public-rdf-dawg-
> comments-request@w3.org] On Behalf Of David Booth
> Sent: Tuesday, September 13, 2011 6:41 PM
> To: public-rdf-dawg-comments
> Cc: eric neumann; Ian Emmons
> Subject: SPARQL Update - Clarifying correct behavior when a graph is
> created implicitly
> 
> The SPARQL Update spec indicates that "If data is inserted into a graph
> that does not exist in the graph store, it *should* be created":
> http://www.w3.org/TR/sparql11-update/#insertData
> However, I've run into an issue with one implementation (Parliament
> 2.7.1) in which the graph *is* created automatically in such
> circumstances, but the graph is not persisted in the store unless it
> was
> created explicitly using a CREATE operation.
> 
> For example, the following query creates the named graph whether line 3
> is commented out or not:
> 
>   PREFIX test: <http://example/test/>
>   DROP SILENT GRAPH test: ;
>   CREATE SILENT GRAPH test: ;   # Line 3
>   INSERT DATA
>   {
>     GRAPH test: { test:foo a test:bar }
>   }
> 
> But if line 3 is commented out, then the graph is not persisted.
> 
> I consider this a bug (and I assume the Parliament implementers will
> also, though I have not heard back from them yet) because it would lead
> to substantial compatibility issues for SPARQL update queries if some
> implementations persisted the graph and others did not.  But AFAICT,
> the
> SPARQL Update spec does not make clear that this behavior is wrong.
> 
> Therefore, I suggest adding some clarification to this effect at the
> end
> of the first paragraph in section 3.1:
> http://www.w3.org/TR/sparql11-update/#graphUpdate
> [[
> If a graph is created implicitly by an update operation, then the
> behavior of the Graph Store MUST be functionally equivalent to its
> behavior if the graph had been created explicitly by a CREATE
> operation.
> ]]
> 
> Thanks!
> 
> 
> --
> David Booth, Ph.D.
> http://dbooth.org/
> 
> Opinions expressed herein are those of the author and do not
> necessarily
> reflect those of his employer.